Jadad Score

The Jadad scale (Jadad et al., 1996) is a simple 0–5 score of RCT quality built from three items: randomisation (up to 2 points), double-blinding (up to 2 points) and a description of withdrawals & dropouts (1 point). A point is added if the method is described and appropriate, and deducted if described but inappropriate. Higher is “better” — but the number hides which item actually scored.

A single headline figure of trial “quality” from 0 to 5. Two of the five available points come from blinding, so the scale leans heavily on it; randomisation contributes two more, and the handling of withdrawals just one. The total tells you a trial scored, say, 3/5 — but not why.

Treat a Jadad score as a rough triage flag, never a verdict. A high score is reassuring only on the three things it measures; a low score should prompt you to look at the individual domains rather than dismiss the trial. Crucially, ask the question Jadad never asks: was allocation concealed?

Jadad illustrates why summary quality scores have fallen out of favour. Collapsing distinct bias mechanisms into one number lets a serious flaw in one domain be masked by strengths in another, and the weighting is arbitrary. Modern appraisal — and the exam — expects domain-by-domain judgement (Cochrane RoB 2), not a single composite score.

• Jadad = randomisation + blinding + withdrawals, max 5
• Blinding is over-weighted; allocation concealment ignored
• Summary scores discouraged → use RoB 2 (domain-based)

Jadad vs RoB 2 on the same trial — an open-label ED analgesia trial with proper computer randomisation but no blinding might score Jadad 3/5 (2 for randomisation, 0 for blinding, 1 for dropouts) — a middling number that says nothing about concealment. Run the same trial through RoB 2 and you get a per-domain picture: low risk for the randomisation process, but high risk for the unblinded subjective outcome. The domain assessment tells you where the problem is; the Jadad number does not. A respectable Jadad score is not a clean bill of health — a trial can score well yet hide an unconcealed allocation that Jadad simply never looks at.

• A summary score hides which domain failed — two trials with the same Jadad total can have very different flaws.
• Forgetting that Jadad ignores allocation concealment, the item Cochrane regards as pivotal.
• Treating Jadad as current best practice — it is outdated; domain-based tools (RoB 2) are preferred.